Black/White Differences in Psychological Adjustment to Spousal …sites.bu.edu › deborahcarr › files › 2018 › 01 › black_widows.pdf · 2018-01-10 · ARTICLERESEARCH ON

10.1177/0164027504268495ARTICLERESEARCH ON AGINGCarr / PSYCHOLOGICAL ADJUSTMENT TO SPOUSAL LOSS

Black/White Differences inPsychological Adjustment to

Spousal Loss Among Older Adults

DEBORAH S. CARRRutgers University

This study examines differences between Blacks and Whites in the effect of widow-hood on depressive symptoms and anxiety; in grief symptoms six months afterspousal loss; and the extent to which these differences are explained by marital qual-ity, social support from children and friends, and religiosity. Analyses are based on theChanging Lives of Older Couples, a prospective study of 1,532 married individualsaged 65 and older. Widowhood is associated with elevated anxiety and depressivesymptoms, yet these effects do not differ by race. Among widowed persons only,Blacks and Whites report similar levels of overall grief, yearning, intrusive thoughts,shock, depressive symptoms, and anxiety, whereas Blacks report significantly lowerlevels of anger and despair. The racial gap in anger is explained by Blacks’higher lev-els of religious participation and social support from children, whereas the differencein despair reflects Blacks’ higher levels of preloss marital conflict.

Keywords: African Americans; bereavement; grief; psychological adjustment;racial differences; widowhood

Widowhood is considered one of the most distressing life events(Holmes and Rahe 1967). The psychological consequences oflate-life widowhood have been documented extensively (see Carr and

591

AUTHOR’S NOTE: An earlier version of this article was presented at the 2003 annual meet-ings of the American Sociological Association. I would like to thank the three anonymous re-viewers for their helpful comments. The Changing Lives of Older Couples (CLOC) has its prin-cipal support from National Institute of Aging (NIA) grants AG15948-01 (Randolph M. Nesse,principal investigator), AG610757-01 (Camille B. Wortman, principal investigator), andAG05561-01 (James S. House, principal investigator). A public-use version of the data is avail-able from the Inter-university Consortium for Political and Social Research at the University ofMichigan or via the Web site http://www.cloc.isr.umich.edu. Address correspondence toDeborah S. Carr, Department of Sociology and Institute for Health, Health Care Policy & AgingResearch, Rutgers University, 30 College Ave., New Brunswick, NJ 08901; e-mail: [email protected].

RESEARCH ON AGING, Vol. 26 No. 6, November 2004 591-622DOI: 10.1177/0164027504268495© 2004 Sage Publications

Utz 2002 for review), yet explorations of racial differences are nearlyabsent from the literature. This omission reflects the fact that few sam-ple surveys include adequate numbers of older Blacks, given their ele-vated risk of premature death (Gibson 1994; Kitson 2000). Studies ofrecently widowed older Blacks are even more difficult, given thatBlacks are less likely than Whites to marry and to remain married overthe life course (Lugaila 1998). Several small qualitative studies havedescribed the experiences of Black and White widows (e.g., Lopata1973), yet few studies have explored systematically racial differencesin older adults’psychological adjustment to spousal loss. Understand-ing the distinctive sources of psychological distress and adjustmentamong Black elders will become increasingly important in the com-ing decades, as Blacks comprise an increasingly large proportion ofthe older population. Although Blacks currently account for 8% of thepopulation over age 65, this proportion will increase to 12% by 2050(U.S. Bureau of the Census 2002).

This study examines (1) whether spousal loss affects depressivesymptoms and anxiety differently for Black and White older adults;(2) whether recently widowed Blacks and Whites experience differentgrief symptomatology; and (3) the extent to which Black-White dif-ferences in grief symptoms are explained by four theoreticallyinformed pathways: marital quality, social support from children,social support from friends and distant relatives, and religiosity.Analyses are based on data from the Changing Lives of Older Couples(CLOC), a prospective study of widowhood among American menand women ages 65 and older.

Theoretical Background

Despite widespread belief that widowhood is among the moststressful life events (Holmes and Rahe 1967), most studies find thatonly 15% to 30% of older adults experience clinically significantdepression in the year following their spouse’s death (Jacobs et al.1989; Lund et al. 1985-1986; Zisook and Shuchter 1991). Less severepsychological reactions are common, however. Depending on thesample and assessment procedures used, an estimated 40% to 70% ofthe recently bereaved experience a period of two or more weeks

592 RESEARCH ON AGING

marked by sadness shortly after the loss (Bruce et al. 1990; Zisooket al. 1997).

At first glance, these statistics suggest that depression and distressare typical reactions to loss, but upon further inspection, they revealthe remarkable psychological resilience of the widowed; at least 70%to 80% experience the widowhood transition without clinical depres-sion, and roughly half survive spousal loss without a two-week periodof sadness or despair. Given that clinical depression is the exception,rather than the norm, in the face of spousal bereavement, researchersface the challenges of identifying the specific psychological symp-toms experienced by the newly bereaved as well as the personal andsocial resources that protect against decrements in psychologicalwell-being among older bereaved spouses. Recent studies have docu-mented that patterns of psychological adjustment to loss vary widelybased on characteristics of the survivor (Matthews 1991; Stroebe andStroebe 1983; Umberson, Wortman, and Kessler 1992), the deceasedspouse (Parkes 1985), the marital relationship (Carr et al. 2000), andthe widowed person’s social relationships (Utz et al. 2002). However,few studies have explored whether Blacks and Whites evidence differ-ent psychological reactions to loss or the ways that racial differencesin social and psychological resources may affect the bereavement pro-cess. Past research on racial differences in marital relationships, socialsupport, and religious coping provides a theoretical foundation forunderstanding Black-White differences in psychological reactions tospousal loss.

MARITAL QUALITY AND ADJUSTMENT TO LOSS

How older adults experience widowhood is linked closely to howthey experienced their late marriages. When a spouse dies, the survi-vor must adjust not only to the loss of an enduring emotional relation-ship but must also manage the daily decisions and household respon-sibilities that were once shared by both spouses (Umberson et al.1992; Utz et al. 2004). A large literature shows that Blacks and Whitesexperience marriage differently and thus may adjust to spousal loss indifferent ways.

First, Black married couples are more likely than their White peersto both endorse and maintain an equitable division of labor within the

Carr / PSYCHOLOGICAL ADJUSTMENT TO SPOUSAL LOSS 593

home. Studies consistently show that Black husbands perform morehours of housework than their White peers and that Black couples areless likely to adhere to a rigid gender-typed division of householdlabor (Beckett and Smith 1981; Dillaway and Broman 2001; Kane1992; Orbuch and Custer 1995; Orbuch and Eyster 1997; Sutherland,Went, and Douvan 1990; Taylor et al. 1991).

The more egalitarian division of household labor among Blackmarried couples has been attributed to long-standing patterns of eco-nomic inequality experienced by Black men. Due to discrimination inthe labor market and educational system, Black men’s educationalattainment, earnings, and job stability lag behind White men’s, andBlack wives historically have worked for pay outside of the home tocontribute to the family’s economic well-being (Hacker 1995; Oliverand Shapiro 1995; Wilson and Neckerman 1986). The division ofhousehold labor tends to be more balanced when wives work outsidethe home for pay, although employed wives still consistently do morehousework than their husbands (Shelton and John 1996).

Black men’s economic marginality has further implications fornegotiations about housework: Black wives may expect (and receive)assistance in housework because their husbands’ economic contribu-tions to the household may be unstable, whereas white husbands may“buy out” of housework with financial contributions to the household(Orbuch and Eyster 1997). Thus, it is possible that widowhood may bea less distressing event for Blacks than for Whites; Black spouses maybe less dependent on one another for the exclusive performance ofimportant gender-typed household tasks and thus are better preparedto manage both household maintenance and homemakingresponsibilities following loss.

Emotional aspects of marriage also differ for Black and Whitespouses. Blacks consistently report lower levels of marital quality andsatisfaction and higher levels of marital conflict than do Whites (e.g.,Acitelli, Douvan, and Veroff 1997; Adelmann, Chadwick, andBaerger 1996; Broman 1993; Goodwin 2003). This gap has beenfound to persist over the life course (e.g., Adelmann et al. 1996; Glenn1989) and net of economic resources and strain (e.g., Broman 1993;Oggins, Veroff, and Leber 1993). These patterns have been attributedto Blacks’ disadvantage in terms of important resources that promotemarital quality, such as good physical health (e.g., Booth and Johnson


1994), financial stability (White and Rogers 2000), and trust in a part-ner’s fidelity (Cazenave and Smith 1990; Lopata 1973).

Racial differences in marital quality may have important implica-tions for how Blacks and Whites adjust to spousal loss. Early psycho-analytic theories of grief proposed that the loss of a conflicted orambivalent marital relationship would be associated with prolongedor “pathological” grief (Abraham [1924] 1927; Freud [1917] 1959).Survivors who had strained relationships with their spouses arebelieved to have both anger toward and a strong attachment to thedeceased. These conflicting feelings make it difficult for survivors tolet go of their loved ones, yet they are also angry at the deceased forabandoning them (Freud [1917] 1959). Recent empirical evaluationsprovide support for an alternative pattern, however; persons with con-flicted relationships report the fewest grief symptoms after theirspouse’s death, whereas those who had the closest marriages experi-ence the most profound grief upon loss (Carr et al. 2000). These find-ings suggest that Blacks may experience fewer grief symptoms thanWhites following spousal loss. Given their higher levels of maritalconflict and lower levels of marital satisfaction than Whites, the lossof a partner may require less profound psychological adjustments.

SOCIAL AND RELIGIOUS SUPPORT

Psychological adjustment to spousal loss also may be affected bythe broader network of social ties maintained by older adults; supportfrom extended family, friends, and participation in formal religiousactivities may be particularly important for older Black bereavedspouses. Some scholars have argued that because Black marriageswere not recognized as legal unions in the United States until afteremancipation, Blacks historically have found social support beyondthe boundaries of the nuclear family or marital dyad and have turnedto extended kinship networks and the church (Genovese 1974; Taylor,Jackson, and Chatters 1997). Recent studies document that Blacks areless likely to depend on and interact with members of the nuclearfamily only and instead maintain a more diffuse social network thatmay include friends, distant relatives, neighbors, and members oftheir church congregation (Ajrouch, Antonucci and Janevic 2001;Chatters, Taylor, and Neighbors 1989; Stack and Burton 1993; Taylor


and Chatters 1986). Blacks also have more frequent contact withmembers of their social networks (Ajrouch et al. 2001) and more fre-quent face-to-face contact with both relatives (Cantor, Brennan, andSainz 1994) and nonrelatives, including church members (Kim andMcKenry 1998). Given that social support is one of the most impor-tant resources for coping with stressful life events (Thoits 1995),Blacks’ more frequent social contacts and more diverse social net-works may provide an important source of instrumental and expres-sive support as they adjust to spousal loss. In this analysis, I examinewhether race differences in psychological adjustment to loss areexplained, in part, by patterns of reliance on two important sources ofnonmarital social support: (1) support from children and (2) supportfrom friends and distant family.

Blacks also are more likely than Whites to participate in formal re-ligious activities and to rely on their religious beliefs as a strategy forcoping with stressful events (Levin, Taylor, and Chatters 1994; Mattisand Jagers 2001). The beneficial effects of religion—particularly forolder adults—have been widely documented (Koenig 1998; Levin,Chatters, and Taylor 1995; Taylor and Chatters 1986, 1991). Onestudy of racial differences in religiosity among older adults found thatBlacks had higher scores than Whites on 19 of 21 possible indicatorsof religiosity (Levin et al. 1994). The importance of religion in theBlack community has been attributed to a history of discriminationand prejudice in the United States; the church historically has pro-vided Blacks with a social and spiritual haven that was built, funded,and controlled by their community (Nelsen and Nelsen 1975).

Religion is a multifaceted construct and encompasses religiousbehavior (such as attendance at services), beliefs (including the use offaith and reliance on God), and social integration (including emo-tional and spiritual support) (Krause 2002). Each of these dimensionsmay provide distinctive benefits to older adults as they cope withspousal loss. First, church attendance may be associated with thereceipt of instrumental, socioemotional, and spiritual support. Per-sons who frequently attend religious services may receive moreinstrumental support than those who do not; frequent interpersonalcontact may make congregants more aware of the needs of olderchurch members (Krause 2002). Second, persons who frequentlyattend services are more likely to engage in social and religious rituals


that enhance solidarity and closeness among congregants, perhaps byincreasing commitment to the ideals, goals, and values of the church(Stark and Finke 2000). Third, persons who regularly attend servicesmay receive spiritual support or assistance that is aimed towardincreasing their religious commitment, beliefs, and behavior. Forinstance, parishioners may share their own religious experiences withrecently widowed older adults or may show them how to apply theirreligious beliefs as they manage their loss (Krause et al. 2001)

Religion also provides a framework for thinking about the worldand for coping with difficult stressors. Religious teachings and beliefsmay provide a sense of certainty (Peterson 2000) and may instill hopeduring times of despair (Levin 2001:138). Persons who have a closerelationship with God may develop a deep sense of trust in God andmay believe that God is in control of their lives, God knows what isbest for them, and God will provide what they need to manage life’schallenges (Koenig 1994). Drawing on past research and theory on theprotective effects of religion, I focus on two aspects of religiosity inthis study: attendance at religious services and religious coping, or theextent to which one’s religious beliefs affect adjustment to, and under-standing of, difficult life stressors. These two aspects of religion mayoffer distinctive types of support to older Blacks and Whites as theyadapt to spousal loss.

OTHER INFLUENCES ONADJUSTMENT TO SPOUSAL LOSS

This analysis includes indicators of three other potential influenceson psychological adjustment to loss. First, I control psychologicalwell-being prior to loss to help distinguish one’s affective state beforethe death and change in affective state that occurred following thedeath (Jacobs 1993; Zisook and Shuchter 1991). Second, I controlboth spouse’s and respondent’s physical health at baseline becausephysical health may affect one’s likelihood of being widowed, as wellas psychological adjustment at the six-month follow-up (Booth andJohnson 1994; Wickrama, Lorenz, and Conger 1997).

Third, I control socioeconomic status prior to loss (includingincome, education, and home ownership) in order to address the pos-sibility that the relationship between spousal loss and psychological


adjustment is spurious. Low socioeconomic status increases one’slikelihood of becoming widowed (Preston and Taubman 1994) and ofexperiencing psychological distress (Miech and Shanahan 2000). It isparticularly important to control socioeconomic status in analyses ofBlack-White differences in psychological adjustment because oflong-standing racial disparities in education and earnings in theUnited States. Blacks are disadvantaged relative to Whites in terms ofeducation, income, and assets (Oliver and Shapiro 1995). Richersocioeconomic resources, particularly education, enable effectivecoping by providing both the financial and psychological resources(such as high levels of perceived control) to manage adversity (Shawand Krause 2001). Studies that fail to control for socioeconomicresources may inaccurately characterize the relationship between raceand psychological adjustment.

In summary, this study will contribute to research on racial differ-ences in late-life spousal bereavement in four ways. First, I examinewhether psychological adjustment of Black and White widowed per-sons differs from still married matched controls. By comparing therecently bereaved with a matched control, I can differentiate the psy-chological effects of widowhood from psychological changes due toaging or the passage of time. Second, I examine whether Black andWhite widowed persons differ in their specific psychological reac-tions to loss by considering a broad array of grief symptoms. Specificgrief symptoms may respond in very different ways to the widowhoodtransition, and these (potentially) competing effects may cancel outone another if only an aggregated scale, such as overall grief, is used asan outcome variable. Third, I examine the extent to whichBlack-White differences in adjustment to loss can be explained byracial differences in how the late marriages were experienced and inother sources of emotional and social support, including religiosityand support from both children and friends. These resources are mea-sured prior to loss and thus are not subject to retrospective recallbias—such as the tendency to retrospectively “sanctify” the memoryof one’s late spouse and marriage (Lopata 1973). Finally, all analysescontrol socioeconomic status and psychological and physical healthprior to the loss in order to address possible confounds in the relation-ships among race, widowhood, and psychological adjustment.


Method

DATA

The CLOC is a prospective study of a two-stage area probabilitysample of 1,532 married individuals from the Detroit standardizedmetropolitan statistical area (SMSA). To be eligible for the study,respondents had to be English-speaking members of a married couplewhere the husband was age 65 or older. All sample members werenoninstitutionalized and were capable of participating in atwo-hour-long interview. Approximately 65% of those contacted foran interview participated, which is consistent with response rates fromother Detroit-area studies. Baseline face-to-face interviews wereconducted in 1987 and 1988.

Spousal loss was monitored by reading the daily obituaries in threeDetroit-area newspapers and by using monthly death record tapes pro-vided by the state of Michigan. The National Death Index (NDI) wasused to confirm deaths and obtain causes of death. Of the 319 respon-dents who lost a spouse during the study period, 86% (n = 276) partici-pated in at least one of the three follow-up interviews that were con-ducted six months (Wave 1), 18 months (Wave 2), and 48 months(Wave 3) after the spouse’s death. Controls from the original baselinesample were selected to match the widowed persons along the dimen-sions of age, race, and sex. The matched controls were reinterviewedat the three follow-up interviews at roughly the same time as thecorresponding widowed persons.

I use two analytic samples in this study. The first includes all 297persons who participated in the six-month follow-up interview. The297 comprise 210 widowed persons (177 White and 33 Black) and 87matched controls (75 White and 12 Black); this sample is used to eval-uate whether the event of widowhood affects psychological adjust-ment differently for Blacks and Whites. The CLOC includes fewercontrols than widowed respondents at the Wave 1 interview becausefunding for data collection was cut from the proposed budget and notreinstated until halfway through the data collection period for Wave 1,thus providing more control respondents for the Wave 2 and 3 inter-views (see Carr and Utz 2002 for further detail on the CLOC study).The second analytic sample includes widowed persons only andallows an exploration of racial differences in how older adults adjust


to spousal loss. The widowed sample includes 210 persons (177White and 33 Black) interviewed at the six-month follow-up.

The issue of selective attrition deserves brief mention. If personswho failed to participate in the six-month follow-up interview are sig-nificantly different from those who did participate (in terms of base-line characteristics), then caution should be taken in generalizing myfindings to the larger population of elderly widowed persons. I esti-mated logistic regression models to identify the correlates ofnonparticipation in the Wave 1 interview. The following variableswere evaluated as possible predictors of attrition: baseline (preloss)demographic and socioeconomic characteristics, marital quality,social support, physical and mental health, and spouse’s health. I alsoseparately evaluated interaction terms of each potential predictor vari-able by race in order to ascertain whether Blacks and Whites have sig-nificantly different sources of sample attrition. Overall, only threevariables were significant predictors of attrition, and these effects didnot differ significantly by race (i.e., race interaction terms were notstatistically significant at the p ≤ .05 level. However, this may reflectthe small sample of Black controls). Age and baseline anxietyincrease the likelihood of nonparticipation, and home ownershipdecreases the likelihood of nonparticipation. Caution should be takenin generalizing findings to the population at large because older, moreanxious and residentially mobile persons may be underrepresented.

MEASURES

Dependent Variables

Two general (i.e., depressive symptoms and anxiety) and sixloss-related (i.e., overall grief, yearning, despair, anger, intrusivethoughts, and shock) dimensions of psychological adjustment at thesix-month follow-up are considered. I focus on the six-month fol-low-up because grief symptoms tend to be most acute during the firstsix months following loss and then decline over time (Zisook andShuchter 1991). Depressive symptoms (α = .83) are assessed with asubset of nine negative items from the 20-item Center for Epidemio-logic Studies–Depression Scale (CES-D; Radloff 1977). Respon-dents indicate how often they experienced each of nine symptoms inthe week prior to interview. Symptoms are the following: I felt


depressed; I felt that everything I did was an effort; My sleep was rest-less; I felt lonely; People were unfriendly; I did not feel like eating.My appetite was poor; I felt sad; I felt that people disliked me; and Icould not get going.

Anxiety (α =. 86) is measured with the Symptom Checklist90–Revised (Derogatis and Cleary 1977). Respondents indicate howoften they experienced each of 10 symptoms in the week prior to inter-view. Response categories were not at all, a little bit, moderately,quite a bit, and extremely. Symptoms were nervousness or shakiness,trembling, feeling suddenly scared for no reason, feeling fearful, heartpounding or racing, feeling tense and keyed up, spells of terror andpanic, feeling so restless you couldn’t sit still, feeling that somethingbad is going to happen to you, and thoughts and images of afrightening nature.

Five specific dimensions of grief also are considered. Yearning (α =.75) was assessed with four questions: In the last month, (1) have youfound yourself longing to have your spouse with you; (2) have you hadpainful waves of missing your spouse; (3) have you experienced feel-ings of intense pain or grief over the loss of your spouse; and (4) haveyou experienced feelings of grief, loneliness, or missing your spouse?Despair (α =. 64) reflects three symptoms felt during the past month:(1) life seemed empty, (2) I felt empty inside, and (3) I felt life had lostits meaning. Anger (α =. 68) is based on three questions: In the pastmonth, have you (1) felt resentful or bitter about the death, (2) felt thedeath was unfair, and (3) felt anger toward God? Intrusive thoughts(α = .66) are based on three symptoms experienced in the past month:(1) difficulty falling asleep because thoughts about your spouse keptcoming into your mind, (2) tried to block out memories or thoughts ofyour spouse, and (3) unable to get thoughts about your spouse out ofyour mind. Shock (α = .77) reflects three symptoms experienced dur-ing the last month: (1) felt in a state of shock, (2) couldn’t believe whatwas happening, and (3) felt emotionally numb. Overall grief (α =.86)is the average of the five subscale scores. Items were drawn fromwidely used grief scales including the Bereavement Index (Jacobs,Kasl, and Ostfeld 1986), Present Feelings About Loss (Singh andRaphael 1981), and Texas Revised Inventory of Grief (Zisook,Devaul, and Click 1982). Dependent variables are standardized forease of interpretation and comparison across indicators. Each scale


has a mean of 0 and standard deviation of 1, where higher scoresreflect more frequent grief symptoms.

Independent Variables

The central independent variable is race, a dichotomous variable setequal to 1 for Blacks. The reference category includes non-HispanicWhites. Four sets of mediating variables are evaluated as pathwaysthat may account for racial differences in psychological adjustment towidowhood: marital quality, social support from children, social sup-port from extended family/friends, and religiosity.

Marital quality. Three characteristics of marriage (evaluated priorto spousal loss) are considered: duration, conflict, and instrumentaldependence. Marital duration is the number of years one had beenmarried to one’s late spouse. Marital conflict (α = .64) is a two-itemscale based on the following items: “How often would you say youand your spouse typically have unpleasant disagreements and con-flicts?” and “In some marriages, there are times when you feel veryclose, but other times when you can get more upset with that personthan with anyone else. How much does this sound like the relationshipyou have with your spouse?” Instrumental dependence is measuredwith the following questions: “Husbands and wives often depend onone another to handle different responsibilities. At the present time,how much do you depend on your spouse to (1) handle or help withhome maintenance and minor repairs; (2) keep up with checking andsavings accounts and pay bills; (3) make major financial and legaldecisions; and (4) prepare meals, general housework, and laundry?”Factor analyses yielded one three-item subscale (α =.54) that tapshome maintenance and financial management tasks that are usuallyperformed by husbands (Items 1 through 3) and a single-item tappinghomemaking tasks, which are typically performed by wives (Item 4).Items evaluating marital conflict and dependence are drawn from theDyadic Adjustment Scale (Spanier 1976).

Religiosity. Two aspects of religiosity are considered: attendance atreligious services and religious coping. Attendance at religious ser-vices is assessed with the question: “How often do you usually attendreligious services? More than once a week, about once a week, one to


three times a month, less than once a month, or never?” Responses arerecoded into two dichotomous variables: never and at least once aweek. The reference category includes persons who attend servicesfewer than three times per month. Religious coping (α = .76) is atwo-item scale based on the following questions: “When you haveproblems or difficulties in your family, work, or personal life, howoften do you seek spiritual comfort and support?” and “When youhave decisions to make in your everyday life, how often do you askyourself what God would want you to do?” Response categories arealmost always, often, sometimes, rarely, or never. The scale is stan-dardized and has a mean of 0 and standard deviation of 1. Higherscores represent higher levels of religious coping.

Social support. Two sources of social support are considered: (1)children and (2) other family members and friends. Dependence onchildren (α = .60) is based on three items: “How much do you dependon your children for emotional support, for help or advice with finan-cial and legal matters, and for help with errands or other chores?” Per-sons who have no living children are assigned the sample mean andare also indicated by a dichotomous variable (1 = has no living chil-dren). Social support from friends and relatives (α = .71) is based onthe following two items: “On the whole, how much do your friendsand relatives make you feel loved and cared for?” and “How much areyour friends and relatives willing to listen when you need to talk aboutyour worries or problems?” Response categories are a lot, some, a lit-tle, or not at all. Both scales are standardized, and higher valuesrepresent greater levels of support.

Well-being at baseline. Preloss indicators of psychological andphysical well-being are controlled to address the possibility that therelationship between widowhood and psychological adjustment isspurious. The characteristics that elevate one’s risk of widowhood,such as poor health, also may be associated with poorer psychologicaladjustment following loss. Depressive symptoms (α = .83) and anxiety(α = .86) are evaluated at baseline with scales identical to those used atthe Wave 1 follow-up (Derogatis and Cleary 1977; Radloff 1977).Respondent’s physical health is assessed with the question: “Howwould you rate your health at the present time? Would you say it isexcellent, very good, good, fair or poor?” Spouse’s physical health (at


baseline) is evaluated with a similar question: “How would you rateyour spouse’s health at the present time?” Both respondent’s andspouse’s health are recoded into dichotomous variables whereresponses of “fair” and “poor” are coded 1 and those with “good” orbetter health are coded 0.

Socioeconomic resources. Socioeconomic resources are controlledto address the possibility that the relationship between widowhoodand psychological adjustment is spurious. Three indicators of socio-economic resources (at baseline) are considered: education (a contin-uous measure ranging from 3 to 17 years of completed schooling),home ownership (1 = owns home), and total household income (natu-ral log of income). Respondents indicate which of 10 income catego-ries most accurately describes their economic status. I derived a con-tinuous measure of income by taking the midpoint of each of the 10income categories, with Pareto estimation of the mean for the topincome category. The natural log of income is used because the distri-bution is skewed, with most respondents in the lower incomecategories.

Demographic variables. The analyses include controls for gender(1 = female), age, and the duration (in months) between the baselineand Wave 1 interviews. All Wave 1 interviews were conducted sixmonths after spousal death, but the duration between the baseline andWave 1 interviews ranges from 9 to 76 months due to variation in thetiming of spouse’s death. Baseline assessments are more temporallydistant for those who lost their spouses at later dates.

ANALYTIC PLAN

The analysis has four parts. First, I present descriptive statistics(means for continuous variables and proportions for dichotomousvariables) and the results of two-tailed t tests comparing values on allvariables for Blacks and Whites in the CLOC sample (Table 1). Sec-ond, I present unadjusted means for Black and White widowed per-sons on each of the grief symptom scales and indicate statistically sig-nificant race differences (Table 2). Third, I use ordinary least squares(OLS) regression models to examine whether the event of widowhoodaffects depressive symptom and anxiety levels among older adults. I


also evaluate race-by-widowhood interaction terms to assess whetherwidowhood affects Blacks and Whites differently (Table 3). Finally, Iuse OLS regression models to examine whether Black and White wid-owed persons differ in their grief symptoms. I evaluate the extent towhich observed race differences are mediated by marital quality,social support from children and friends, and religiosity (Tables 4and 5).

Results

SAMPLE CHARACTERISTICS

Descriptive statistics and t tests comparing means for Blacks andWhites are presented in Table 1. Asterisks denote significantBlack-White differences within the total (column 1), control (column2), and widowed samples (column 3). Overall, Blacks have signifi-cantly lower levels of income and education than Whites. Blacks andWhites also differ in terms of marital conflict, religiosity, and depend-ence on their children. No racial differences are found for prelosshealth and well-being, instrumental dependence on one’s spouse, andsupport from friends and other family members.

The CLOC data reveal pronounced racial differences in how mar-riage is experienced. In the total sample (column 1), Blacks report lev-els of marital conflict at baseline that are roughly one-half standarddeviation higher than that of Whites (.33 versus –.21, p < .001) andsignificantly shorter marriages (37 versus 42 years, p < .05). However,Blacks are advantaged in terms of two other psychosocial resources:religiosity and support from children. In the total sample, Blacks havesignificantly higher levels of religious coping (.66 versus –.12, p <.001), are more likely to attend religious services at least weekly (70%vs. 51%, p < .05), and are less likely to report that they “never” attendservices (2% versus 19%, p < .01). They also report greater reliance ontheir children for instrumental and emotional support (.59 versus .06,p < .001).

Blacks and Whites do not differ in terms of depressive symptoms oranxiety at either the baseline interview or six-month follow-up.Although the gap in depressive symptoms between the nonwidowedand widowed is almost twice as large among Whites compared to


606

TAB

LE

1

Mea

ns a

nd S

tand

ard

Dev

iatio

ns f

or B

lack

s an

d W

hite

s, C

hang

ing

Liv

es o

f O

lder

Cou

ples

(N

= 2

97)

Tota

l Sam

ple

Con

trol

Sam

ple

Ber

eave

d Sa

mpl

e(N

= 2

97)

(n=

87)

(n=

210

)

Bla

cks

Whi

tes

Bla

cks

Whi

tes

Bla

cks

Whi

tes

MSD

MSD

MSD

MSD

MSD

MSD

Dep

ende

nt v

aria

bles

Dep

ress

ive

sym

ptom

s (C

ES-

D)a

0.16

51.

270.

269

1.16

–0.0

631.

07–0

.156

0.97

80.

249

1.34

0.44

61.

19A

nxie

ty (

SCL

-90)

a–0

.189

0.57

50.

034

1.06

–0.3

610.

399

–0.0

771.

03–0

.125

0.62

20.

079

1.07

SES

and

dem

ogra

phic

cha

ract

eris

tics

Edu

catio

n10

.68

3.02

11.5

12.

8411

.86

3.25

1.64

2.73

10.2

42.

8611

.46*

2.89

Inco

me,

bas

elin

e15

,507

11,1

3422

,804

**17

,138

17,6

5210

,182

24,2

1917

,652

14,7

1111

,515

22,2

10*

16,9

33In

com

e (n

atur

al lo

g)1.

120.

490

1.39

**0.

501

1.29

0.37

31.

450.

475

1.06

0.51

71.

36**

0.51

1O

wns

hom

e0.

886

0.32

20.

939

0.23

80.

916

0.29

00.

976

0.15

60.

870.

340.

920.

26Se

x (1

= f

emal

e)0.

772

0.42

40.

723

0.44

90.

916

0.29

00.

728

0.44

80.

720.

460.

720.

45A

ge69

.48

7.55

70.2

36.

5866

.99

6.93

69.3

85.

9170

.47.

6670

.66.

82

Pre

loss

hea

lth

and

wel

l-be

ing

Dep

ress

ive

sym

ptom

s (C

ES-

D)

0.05

90.

978

0.03

91.

02–0

.065

0.75

8–0

.009

1.15

0.10

41.

060.

060.

963

at b

asel

inea

Anx

iety

at b

asel

inea

0.08

90.

917

–0.0

560.

717

0.06

10.

802

–0.1

590.

717

0.09

90.

967

–0.0

130.

714

Fair

/poo

r ph

ysic

al h

ealth

at b

asel

ine

0.30

60.

466

0.30

70.

462

0.30

10.

479

0.24

40.

433

0.30

80.

469

0.33

40.

473

Spou

se in

fair

/poo

r he

alth

at b

asel

ine

0.54

40.

504

0.50

20.

501

0.39

50.

510

0.26

90.

447

0.59

90.

498

0.60

10.

491

607

Mar

ital

qua

lity

Mar

ital d

urat

ion

in y

ears

(ba

selin

e)37

.38

13.9

742

.10*

13.2

934

.82

12.0

138

.53

15.1

338

.33

14.6

943

.61*

12.1

8M

arita

l con

flic

t at b

asel

inea

0.33

31.

17–0

.210

***

0.96

10.

585

1.33

0.77

7–0

.256

***

0.23

91.

12–0

.073

*0.

914

Dep

ende

nce,

hom

emak

ing

task

sa0.

088

0.95

8–0

.016

1.01

–0.3

830.

723

–0.2

211.

020.

262

0.98

50.

071

0.99

5D

epen

denc

e, h

ome

mai

nten

ance

0.17

10.

656

–.02

00.

806

0.56

10.

517

0.33

10.

647

0.02

7a0.

649

–0.1

670.

822

and

fina

ncia

l tas

ksa

Rel

igio

sity

Nev

er a

ttend

s re

ligio

us s

ervi

ces

0.02

0.14

90.

193*

**0.

395

0.04

20.

209

0.16

20.

371

0.01

50.

123

0.20

6**

0.40

5A

ttend

s re

ligio

us s

ervi

ces

at le

ast

0.70

20.

463

0.50

8***

0.50

10.

666

0.49

20.

535

0.50

20.

715

0.45

90.

496*

0.50

1on

ce p

er w

eek

Rel

igio

us c

opin

g sc

alea

0.65

80.

702

–0.1

16**

*1.

000.

810.

75–0

.069

**1.

030.

602

0.68

7–0

.136

***

0.99

1

Soci

al s

uppo

rt

Dep

ende

nce

on c

hild

ren

at b

asel

inea

0.59

31.

190.

065*

**0.

908

0.81

91.

29–0

.036

**0.

900

0.50

91.

160.

107

0.91

0H

as n

o liv

ing

child

ren

0.12

60.

335

0.07

60.

266

00.

000

0.10

30.

306

0.17

20.

384

0.06

5*0.

247

Ove

rall

supp

ort f

rom

fri

ends

/rel

ativ

es0.

227

0.93

20.

044

1.03

0.56

80.

632

0.05

91.

040.

101

1.00

0.03

71.

03at

bas

elin

ea

Mon

ths

betw

een

base

line

and

41.4

918

.59

43.3

619

.37

60.9

10.4

859

.16

7.96

34.2

915

.54

36.7

218

.90

Wav

e 1

inte

rvie

wN

4525

212

7533

177

NO

TE

:W

eigh

ted

data

are

repo

rted

inth

eta

ble.

Two-

taile

dt

test

sw

ere

used

toas

sess

sign

ific

ant

diff

eren

ces

betw

een

mea

ns.

Ast

eris

ksde

note

sign

ific

ant

Bla

ck-W

hite

diff

eren

ces

inth

eto

tal(

colu

mn

1),c

ontr

ol(c

olum

n2)

,and

wid

owed

sam

ples

(col

umn

3).C

ES-

D=

Cen

ter

for

Epi

dem

iolo

gic

Stud

ies–

Dep

ress

ion

Scal

e; S

CL

-90

= S

ympt

om C

heck

list 9

0–R

evis

ed; S

ES

= s

ocio

eono

mic

sta

tus.

a.T

hese

var

iabl

es a

re s

tand

ardi

zed.

*p<

.05.

**p

< .0

1. *

**p

< .0

01.

Blacks, the difference is not statistically significant. Blacks andWhites also report similar levels of most grief symptoms. Table 2presents Blacks’and Whites’ (unadjusted) mean levels of grief symp-toms, depressive symptoms, and anxiety, among the 210 widowedpersons interviewed six months following spousal loss; standardizedscores are shown. Whites have significantly higher levels of despair(.095 versus –.52) and anger (.076 versus –.41) than do Blacks. Whiteand Black widowed elders do not differ in terms of overall grief,yearning, intrusive thoughts, shock, depression, or anxiety.

MULTIVARIATE ANALYSES

Black-White Differences in Effect of Widowhood

The first objective of the multivariate analysis is to investigatewhether widowhood has significantly different effects on the depres-sive symptoms and anxiety levels of Blacks and Whites. Table 3 dis-plays OLS regression models evaluating the main effects of widow-


TABLE 2

Means and Standard Deviations for Grief, Anxiety,and Depressive Symptoms, Among Black and White Bereaved

Spouses, Six Months Following Spousal Loss (N = 210)

Unadjusted Means

Blacks Whites(n = 33) (n = 177)

M SD M SD

Overall grief –0.30 0.99 0.06 0.99Yearning –0.30 1.05 0.06 0.98Despair –0.52 0.71 0.095*** 1.02Anger –0.41 0.74 0.076* 1.02Intrusive thoughts 0.10 1.03 –0.02 0.99Shock –0.05 0.93 0.01 1.01Depressive symptoms (CES-D) 0.25 1.34 0.45 1.2Anxiety (SCL-90) –0.13 0.62 0.08 1.07

NOTE: Weighted data are reported in the table. All subscales are standardized, where M = 0 andSD = 1. Two-tailed t tests were used to compare unadjusted mean scores for Black and White be-reaved older spouses. CES-D = Center for Epidemiologic Studies–Depression Scale; SCL-90 =Symptom Checklist 90–Revised.*p < .05. **p < .01. ***p < .001.


TABLE 3

Ordinary Least Squares Regression of DepressiveSymptoms and Anxiety (at six-month follow-up)

on Widowhood Status, Race, and Mediating Variables,Changing Lives of Older Couples Study (N = 297)

Depressive Symptoms Anxiety

Model 1 Model 2 Model 1 Model 2

Demographic characteristics

Widowhood (1 = widow) .78*** .82*** .32* .30*(.17) (.18) (.14) (.15)

Race (1 = Black) –.10 .11 –.22 –.31(.18) (.34) (.15) (.29)

Widowhood × Black –.30 .14(.40) (.34)

Age .01 .01 –.01 –.01(.01) (.01) (.01) (.01)

Sex (1 = female) .11 .11 –.01 –.01(.15) (.15) (.13) (.13)

Socioeconomic resources

Years of education .01 .01 .02 .03(.02) (.02) (.02) (.02)

Own home, baseline .29 .29 .26 .27(.25) (.25) (.22) (.22)

Income (natural log), baseline –.07 –.07 –.04 –.04(.14) (.14) (.11) (.11)

Baseline well-being

Depressive symptoms (CES-D) .36*** .36*** .28*** .28***at baseline (.07) (.07) (.06) (.06)

Anxiety, baseline –.01 –.01 .1 .1(.10) (.10) (.08) (.08)

Self-rated health fair or poor .25 .25 .42*** .42***at baseline (.15) (.15) (.12) (.12)

Spouse’s health fair or poor .15 -.15 –.25* –.25*at baseline (.14) (.14) (.15) (.15)

Adjusted R2 .15 .15 .16 .16

Constant –1.94 –1.95 –.04 –.04(.92) (.92) (.77) (.77)

NOTE: Standardized regression coefficients and standard deviations (in parentheses) are shown.Dependent variables are standardized, with mean of 0 and standard deviation of 1. All modelscontrol for number of months between baseline and follow-up interview. CES-D = Center forEpidemiologic Studies–Depression Scale.*p < .05. **p < .01. ***p < .001.

hood and race on the two outcomes (Model 1) as well as a two-wayinteraction term between widowhood and race (Model 2). A statisti-cally significant interaction term indicates that widowhood affectsBlacks and Whites differently. The analyses show that race is not asignificant predictor of either depressive symptoms or anxiety.Widowhood is a significant predictor of both depressive symptoms(b = .8, p < .001) and anxiety (b = –.3, p < .05) at the six-month fol-low-up, yet these relationships do not differ significantly by race.

The lack of statistical significance in the interaction term analysescould reflect low statistical power, given that the widowed-controlsample includes only 12 Black control respondents at the 6-month fol-low-up. To further investigate the possibility that small sample size(rather than the actual absence of significant race differences)accounts for these findings, I reestimated the models shown in Table3, using the widowed-control sample at the 18-month follow-up. Thissample includes 168 widowed persons (144 Whites and 24 Blacks)and 202 controls (170 Whites and 32 Blacks). The CLOC data collec-tion team obtained more interviews among controls than widowedpersons at Wave 2, so that the total number of widowed persons andcontrols combined across all interview waves would be roughly equal.The replicated analysis on the larger Wave 2 sample also revealed thatwidowhood is associated with a significant increase in depressivesymptoms but has no significant effect on anxiety (net of baselinehealth, demographic, and socioeconomic characteristics). Theseeffects do not differ significantly by race, nor is race a significant pre-dictor of either depressive symptoms or anxiety. (Tables are availablefrom the author.)

Race Differences in Grief Symptomatology

The next two objectives of the multivariate analysis are to (1) exam-ine racial differences in grief symptoms six months following spousaldeath and (2) evaluate the extent to which these differences are attrib-utable to differences in Blacks’and Whites’marital quality and socialsupport resources. To investigate the first objective, I estimated OLSregression models predicting each of the five grief symptoms and theoverall grief scale. A series of stepwise models was estimated; thebaseline models included race and demographic characteristics only,


whereas subsequent models incorporated socioeconomic status indi-cators, preloss well-being characteristics, and each of the four poten-tial mediator variables. Neither the gross nor net effects of race werestatistically significant for four of six outcomes: overall grief, yearn-ing, intrusive thoughts, and shock (results not shown, but are availablefrom author). The remaining two symptoms, anger and despair, aresignificantly lower among Blacks than Whites (results shown inTables 4 and 5, respectively).

Race Differences in Anger Six Months After Loss

Table 4 presents OLS regression models evaluating the predictorsof anger six months after spousal loss. Black bereaved spouses reportsignificantly lower levels of anger six months following their spouse’sdeath, and this racial gap declines by roughly 20% when religious par-ticipation and coping are controlled. In Model 1, where only demo-graphic, socioeconomic status, and baseline health factors are con-trolled, Blacks evidence levels of anger that are one-half standarddeviation lower than that of Whites. The effect of race on angerdeclines by roughly 10% (i.e., from –.50 to –.44) when religious par-ticipation is controlled (Model 2) and declines by 20% (i.e., from –.50to –.39) when religious coping is controlled (Model 3). However,when both forms of religiosity are controlled in Model 4, frequentchurch attendance only remains a significant predictor of anger (b =–.47), and the racial gap in anger equals roughly .4 standard devia-tions. Reliance on children for social support (Model 5) is associatedwith reduced levels of anger (b = –.15), and the racial gap in anger isno longer statistically significant when both support from childrenand religiosity are controlled.

Marital characteristics (i.e., duration, conflict, and dependence) areneither significant predictors of widowed persons’ anger nor media-tors of the relationship between race and anger. Neither socio-economic resources nor preloss well-being are significant correlatesof anger. Age is related inversely to anger; each additional year of ageis associated with a .03 standard deviation decrease in anger symp-toms. Gender is associated with anger in Model 1; women reportanger symptoms that are .35 standard deviations lower than men.



TABLE 4

Ordinary Least Squares Regression of Anger(at six-month follow-up) on Race and Mediating Variables,

Changing Lives of Older Couples Study (N = 210)

Model 1 Model 2 Model 3 Model 4 Model 5


Race (1 = Black) –.50** –.44* –.39* –.39* –.36(.19) (.19) (.19) (.19) (.19)

Age –.031** –.029** –.029** –.028** –.029**(.010) (.010) (.010) (.010) (.010)

Sex (1 = female) –.353* –.272 –.284 –.226 –.255(.157) (.155) (.159) (.159) (.159)


Years of education .025 .026 .019 .021 .02(.025) (.024) (.025) (.025) (.024)

Own home, baseline .142 .233 .159 .241 .214(.247) (.243) (.245) (.243) (.241)

Income (natural log), baseline –.065 –.099 –.059 –.098 –.110(.144) (.141) (.142) (.141) (.139)

Baseline well-being

Depressive symptoms (CES-D), .137 .119 .105 .102 .105baseline (.080) (.079) (.081) (.079) (.079)

Anxiety, baseline .072 .087 .115 .109 .117(.078) (.077) (.079) (.078) (.078)

Self-rated health fair or poor –.107 –.149 –.103 –.138 –.148at baseline (.152) (.149) (.151) (.149) (.148)

Spouse’s health fair or poor –.141 –.180 –.165 –.192 –.186at baseline (.142) (.139) (.142) (.139) (.139)

Religiosity

Never attends religious services –.262 –.351 –.336(.201) (.211) (.211)

Attends religious services at least –.527*** –.469** –.437**one per week (.154) (.159) (.159)

Religious coping scale (standardized) –.156* –.113 –.107(.074) (.085) (.209)

Filial support

Dependence on children for support –.146*and assistance (.068)

Has no living children .113(.242)

Adjusted R2 .085 .129 .101 .132 .145Constant 2.34 2.4 2.17 2.31 2.45

(.910) (.889) (.905) (.891) (.889)



TABLE 5

Ordinary Least Squares Regression of Despair(at six-month follow-up) on Race and Mediating Variables,

Changing Lives of Older Couples Study (N = 210)

Model 1 Model 2 Model 3 Model 4


Race (1 = Black) –.543** –.468* –.544** –.459*(.191) (.194) (.189) (.192)

Age –.001 –.015 0.003 –.013(.010) (.013) (.010) (.013)

Sex (1 = female) –.011 –.056 .036 –.019(.159) (.161) (.183) (.184)


Years of education .039 .032 .044 .037(.025) (.025) (.025) (.025)

Own home, baseline .207 .228 .141 .159(.249) (.249) (.248) (.246)

Income (natural log), baseline .025 .024 .059 .063(.145) (.144) (.144) (.142)

Baseline well-being

Depressive symptoms (CES-D) at baseline .104 .095 .162* .158*(.081) (.081) (.083) (.082)

Anxiety, baseline .091 .153 .051 .119(.105) (.108) (.104) (.107)

Self-rated health fair or poor at baseline .165 .155 .157 .146(.154) (.153) (.154) (.152)

Spouse’s health fair or poor at baseline .116 .098 .238 .231(.144) (.143) (.149) (.147)

Marital relationship

Years married .006 .007(.007) (.007)

Marital conflict –.134* –.157*(.071) (.070)

Dependence, home maintenance andfinancial tasks .169* .193*

(.088) (.087)

Dependence, homemaking tasks –.134 –.145*(.075) (.074)

Adjusted R2 .065 .08 .09 .112

Constant –.858 –.088 –1.09 –.221(.919) (.976) (.912) (.959)


Black-White Differences in Despair Six Months After Loss

Table 5 presents OLS regression models evaluating the predictorsof despair six months after spousal loss. Blacks have levels of despairthat are one-half standard deviation lower than Whites (b = –.54),when demographic, socioeconomic, and baseline health characteris-tics are controlled (Model 1). Not one of the socioeconomic status orhealth variables is significantly linked to despair, and only 7% of thevariance in despair is explained by the baseline variables. Each of themediator variables was evaluated, and only the marital quality charac-teristics were significant predictors of despair.

Model 2 adjusts for marital conflict and marital duration; the inclu-sion of these variables results in a 10% decline in the racial differencein despair. Moreover, higher levels of marital conflict at baseline arelinked to lower levels of despair, or one’s sense of emotional empti-ness following spousal loss. Model 3 incorporates indicators of instru-mental dependence in the marriage; although these variables do notmediate the effect of race, one indicator of instrumental dependence issignificantly linked to despair. Persons who were dependent on theirspouse for home repair and financial matters have higher levels ofdespair (b = .17, p < .05), whereas those who were dependent forhomemaking tasks have lower levels of despair (b = –.13, p < .10).Even after martial dependence, duration, and conflict are controlled,however, the race gap in despair persists (b = –.46, p < .05). This find-ing highlights the need for further research on the ways that Blacksand Whites adapt emotionally to the death of their spouses.

Discussion

The analyses revealed few Black-White differences in the waysthat older adults adjust psychologically to the death of a spouse. Com-pared with married matched controls, the widowed evidenced signifi-cantly higher levels of anxiety and depressive symptoms, but thesepatterns do not differ by race. When widowed persons only were con-sidered, Blacks andWhites did not differ in terms of four grief symp-toms: overall grief, yearning, shock, and intrusive thoughts. Blackshad significantly lower levels of two grief symptoms: despair andanger.


Blacks’ lower levels of anger are attributable partly to their higherlevels of religious participation and coping and also are explained par-tially by Blacks’ greater reliance on their children for social support.These findings are consistent with the argument that anger is symp-tomatic of the most socially isolated widowed (Parkes and Weiss1983). Blacks’ lower levels of anger are due in part to their social inte-gration with the religious community and their children, underscoringthe importance of having a broad and varied base of social support.Widowhood is believed to be particularly distressing in individualisticsocieties where the nuclear family is socially and economically auton-omous and where spouses may have few alternative sources of social,emotional, or instrumental support (Lopata 1973; Volkart andMichael 1957). African Americans, perhaps due to their lower rates ofmarriage and greater likelihood of divorce, have developed a strongerand more varied web of social relationships, including extended fam-ily and the church; it is precisely these resources that may bufferagainst symptoms such as anger in the face of spousal loss.

Blacks’ lower levels of despair are explained, in part, by the factthat they report higher levels of marital conflict than do Whites.Although early psychoanalytic perspectives on grief proposed that theloss of a conflicted relationship is associated with prolonged or“pathological” grief and longing (Abraham [1924] 1927; Freud[1917] 1959), more recent empirical analyses reveal that widowedpersons confronting the loss of conflicted or ambiguous relationshipstend to grieve less for their deceased partners, whereas those in closeloving relations yearned most for the deceased (Carr et al. 2000).Thus, Blacks may experience less despair and emotional emptinessbecause they are losing a less emotionally rewarding relationship.Future research should explore whether the linkage between maritalconflict and adjustment to loss differs for Blacks and Whites. Recentresearch suggests that Black and White couples have different expec-tations and criteria for evaluating their marriages (Acitelli et al. 1997;Chadiha, Veroff, and Leber 1998); if Blacks are more likely thanWhites to both anticipate and acknowledge marital conflict, then per-haps the inverse relationship between marital conflict and grief maybe weaker for Blacks than for Whites.

Although I expected that Blacks would evidence lower levels ofdependence on their spouses for homemaking and home maintenancetasks and consequently, lower levels of despair, the CLOC data did not


support this hypothesis. There are several possible reasons why theCLOC data did not confirm the widely documented observation thatBlack and White spouses adhere to more egalitarian gender roles inthe family (e.g., Dillaway and Broman 2001; Orbuch and Eyster1997). First, long-standing patterns of allocating household responsi-bilities may change in later life; declines in physical health may meanthat older adults give up tasks that they can no longer perform, thusleaving their spouses to take on new responsibilities even before wid-owhood (Szinovacz 2000; Szinovacz and Harpster 1994). Second, thesmall sample size prevented sex- and race-specific analyses; it is pos-sible that dependence on one’s spouse for “male-typed” tasks is asso-ciated with women’s distress only and vice versa. Future studiesshould evaluate the extent to which a sex-typed or egalitarian divisionof household labor among married couples affects overall adjustmentto spousal loss and racial differences in adjustment to spousal loss.

Overall, the study findings have potentially important implicationsfor understanding psychological adjustment among older bereavedspouses. First, psychological reactions to spousal death appear to beclosely tied to social patterns established earlier in the life course,such as marital relationships, religious participation, and interactionswith one’s children. For instance, the racial gap in anger symptomsattenuated when preloss levels of religiosity and parent-child depend-ence were controlled, suggesting that enduring patterns of roles andrelationships may be important resources for coping with loss. How-ever, I did not explore the extent to which these social roles and rela-tionships change in the face of loss. Future research should explorewhether Blacks and Whites alter their religious beliefs or behaviorfollowing loss and the extent to which reliance on other familymembers changes following the loss of spouse.

Second, these results underscore the importance of consideringmultiple psychological outcomes when studying psychologicaladjustment to loss. The data revealed racial differences in levels ofanger and despair, but not in more global mental health outcomes suchas depressive symptoms, anxiety, or overall grief. If only global indi-cators—rather than precise grief symptoms—had been considered,then the Black-White differences in anger and despair would havegone undetected. Moreover, the consideration of a diverse array ofgrief symptoms allows researchers to move away from the question“who suffers worse” in the face of loss and to instead identify the spe-


cific psychological reactions experienced by distinctive social groupsand demographic categories (e.g., Stroebe and Stroebe 1983).

LIMITATIONS AND FUTURE DIRECTIONS

This study has several important limitations. First, by design, theCLOC sample includes the most physically and economically advan-taged, given that all sample members were married and were age 65 orolder at the time of the baseline interview. Blacks tend to have higherlevels of mortality (Gibson 1994) and lower rates of both marryingand staying married than Whites (Lugaila 1998), therefore the Blacksample may be more selective (and less representative) than theirWhite peers. This potential positive selection bias also may contributeto the slightly better psychological adjustment evidenced by Blacks inthe CLOC sample.

Second, the small sample size prevented the analysis of morefine-grained racial differences, such as within-race gender or socio-economic status differences in adjustment to spousal loss. Examiningonly broad Black-White differences implies that between-race differ-ences are larger or substantively more important than within-racesources of variation in psychological adjustment to loss. Identifyingthe latter is an important pursuit for developing both racially sensitiveand individually targeted programs and interventions to help olderadults adjust to spousal loss (Alvidrez, Azocar, and Miranda 1996;Phinney 1996). Moreover, because of the small sample size and lowstatistical power, few relationships were statistically significant.Additional research, based on larger samples, must be conductedbefore strong conclusions can be drawn about racial differences inadjustment to late life spousal loss.

Third, all spousal deaths were treated similarly in the analysis; theassumption was that the effects of widowhood are invariant regardlessof the cause, timing, or context of the death. However, the psychologi-cal consequences of widowhood have been found to vary based on theduration of the late spouse’s illness and the amount of forewarningone had (Carr et al. 2001). The context and cause of death also matter;deaths that are due to medical negligence or where the dying patientwas in severe pain are much more distressing to survivors (Carr 2003).Given well-documented racial differences in the cause and timing ofdeath (National Center for Health Statistics 2000) and evidence that


Blacks receive poorer quality health care than do Whites (Livingston1994; Williams and Collins 1995), it is important also to explore howthe late spouse’s dying process affects the psychological adjustmentof Black and White older adults.

Despite these weaknesses, the analyses document the differentgrief symptoms experienced by Black and White widowed olderadults and suggest ways that race differences in marital relations andsocial support, broadly defined, account for Black-White differencesin adjustment to loss. This study should be regarded as an importantpreliminary look and springboard for future research, rather than as adefinitive statement about Black and White widowed older adults.

REFERENCES

Abraham, Karl. [1924] 1927. “A Short Study of the Development of the Libido, Viewed in theLight of Mental Disorder.” Pp. 418-50 in Selected Papers of Karl Abraham. London:Hogarth.

Acitelli, Linda, K., Elizabeth Douvan, and Joseph Veroff. 1997. “The Changing Influence ofInterpersonal Perceptions on Marital Well-Being Among Black and White Couples.” Jour-nal of Social and Personal Relationships 14:291-304.

Adelmann, Pamela K., Kirsten Chadwick, and Dana R. Baerger. 1996. “Marital Quality of Blackand White Adults Over the Life Course.” Journal of Social and Personal Relationships13:361-84.

Ajrouch, Kristine, Toni C. Antonucci, and Mary R. Janevic. 2001. “Social Networks AmongBlacks and Whites: The Interaction Between Race and Age.” Journal of Gerontology: SocialSciences 56B:S112-118.

Alvidrez, Jennifer, Francisca Azocar, and Jeanne Miranda. 1996. “Demystifying the Concept ofEthnicity for Psychotherapy Researchers.” Journal of Consulting and Clinical Psychology64:903-8.

Beckett, Joyce O. and Audrey M. Smith. 1981. “Work and Family Roles: Egalitarian Marriage inBlack and White Families.” Social Service Review 55:314-26.

Booth, Alan and David R. Johnson. 1994. “Declining Health and Marital Quality.” Journal ofMarriage and the Family 56:218-23.

Broman, Clifford L. 1993. “Race Differences in Marital Well-Being.” Journal of Marriage andFamily 55:724-32.

Bruce, Martha L., Kathleen Kim, Philip J. Leaf, and Selby Jacobs. 1990. “Depressive Episodesand Dysphoria Resulting From Conjugal Bereavement in a Prospective Community Sam-ple.” American Journal of Psychiatry 147:608-11.

Cantor, Marjorie, Mark Brennan, and Anthony Sainz. 1994. “The Importance of Ethnicity in theSocial Support Systems of Older New Yorkers: A Longitudinal Perspective (1970-1990).”Journal of Gerontological Social Work 22:95-128.

Carr, Deborah. 2003. “A ‘Good Death’for Whom? Quality of Spouse’s Death and PsychologicalDistress Among Older Widowed Persons.” Journal of Health and Social Behavior44:215-32.


Carr, Deborah, James S. House, Ronald C. Kessler, Randolph M. Nesse, John Sonnega, andCamille B. Wortman. 2000. “Marital Quality and Psychological Adjustment to WidowhoodAmong Older Adults: A Longitudinal Analysis.” Journal of Gerontology: Social Sciences55:S197-S207.

Carr, Deborah, James S. House, Camille B. Wortman, Randolph M. Nesse, and Ronald C.Kessler. 2001. “Psychological Adjustment to Sudden and Anticipated Spousal Death Amongthe Older Widowed.” Journal of Gerontology: Social Sciences 56B:S237-S248.

Carr, Deborah and Rebecca L. Utz. 2002. “Late-Life Widowhood in the United States: NewDirections in Research and Theory.” Ageing International 27:65-88.

Cazenave, Noel A. and Rita Smith. 1990. “Gender Differences in the Perception of BlackMale-Female Relationships and Stereotypes.” Pp. 149-70 in Black Families: Interdisciplin-ary Perspectives, edited by Harold E. Cheatham and James B. Stuart. New Brunswick, NJ:Transaction Books.

Chadiha, Letha A., Joseph Veroff, and Douglas Leber. 1998. “Newlyweds’ Narrative Themes:Meaning in the First Year of Marriage for African American and White Couples.” Journal ofComparative Family Studies 29:115-30.

Chatters, Linda M., Robert J. Taylor, and Harold W. Neighbors. 1989. “Size of Informal HelperNetwork Mobilized During a Serious Personal Problem Among Black Americans.” Journalof Marriage and the Family 51:667-76.

Derogatis, Leonard R. and Paul A. Cleary. 1977. “Confirmation of the Dimensional Structure ofthe SCL-90: A Study in Construct Validation.” Journal of Clinical Psychology 33:981-89.

Dillaway, Heather and Clifford Broman. 2001. “Race, Class and Gender in Marital Satisfactionand Division of Household Labor Among Dual-Earner Couples.” Journal of Family Issues22:309-27.

Freud, Sigmund. [1917] 1959. “Mourning and Melancholia.” Pp. 239-58 in The Standard Editionof the Complete Psychological Works of Sigmund Freud, vol. 14, edited by James Strachey.London: Hogarth.

Genovese, Eugene. 1974. Roll, Jordan, Roll: The World the Slaves Made. New York: Pantheon.Gibson, Rose. 1994. “The Age-by-Race Gap in Health and Mortality in the Older Population: A

Social Science Research Agenda.” The Gerontologist 34:454-62.Glenn, Norval D. 1989. “Duration of Marriage, Family Composition, and Marital Happiness.

National Journal of Sociology 3:3-24.Goodwin, Paula Y. 2003. “African American and European American Women’s Marital

Well-Being.” Journal of Marriage and Family 65:550-60.Hacker, Andrew. 1995. Black and White: Separate, Unequal, Hostile. New York: Scribner.Holmes, Thomas H. and Richard H. Rahe. 1967. “The Social Readjustment Scale.” Journal of

Psychosomatic Research 11:213-28.Jacobs, Selby. 1993. Pathologic Grief: Maladaptation to Loss. Washington, DC: American Psy-

chiatric Press.Jacobs, Selby, F. Hansen, Lisa Berkman, Stanislav V. Kasl, and Adrian Ostfeld. 1989. “Depres-

sions of Bereavement.” Comprehensive Psychiatry 31:218-24.Jacobs, Selby, Stanislav V. Kasl, and Adrian Ostfeld. 1986. “The Measurement of Grief:

Bereaved Versus Non-Bereaved.” The Hospice Journal 2:21-36.Kane, Emily W. 1992. “Race, Gender and Attitudes Towards Gender Stratification.” Social Psy-

chology Quarterly 55:311-20.Kim, Hyoun K. and Patrick C. McKenry. 1998. “Social Networks and Support: A Comparison of

African Americans, Asian Americans, Caucasians, and Hispanics.” Journal of ComparativeFamily Studies 29:313-34.

Kitson, Gay C. 2000. “Adjustment to Violent and Natural Deaths in Later and Earlier Life forBlack and White Widows.” Journal of Gerontology: Social Sciences 55B:341-51.


Koenig, Harold G. 1994. Aging and God: Spiritual Pathways to Mental Health in Midlife andLater Years. Binghamton, NY: Haworth.

. 1998. “Religious Beliefs and Practices of Hospitalized Medically Ill Older Adults.”International Journal of Geriatric Psychiatry 13:213-24.

Krause, Neal. 2002. “Church-Based Social Support and Health in Old Age: Exploring Variationsby Race.” Journal of Gerontology: Social Sciences 57B:332-47.

Krause, Neal, Christopher G. Ellison, Benjamin A. Shaw, Jack P. Marcum, and Jason D.Boardman. 2001. “Church-Based Social Support and Religious Coping.” Journal for the Sci-entific Study of Religion 40:653-78.

Levin, Jeffrey S. 2001. God, Faith and Health: Exploring the Spirituality-Healing Connection.New York: John Wiley.

Levin, Jeffrey S., Linda M. Chatters, and Robert J. Taylor. 1995. “Religious Effects on HealthStatus and Life Satisfaction Among Black Americans.” Journal of Gerontology: Social Sci-ences 50B:S154-63.

Levin, Jeffrey S., Robert J. Taylor, and Linda M. Chatters. 1994. “Race and Gender Differencesin Religiosity Among Older Adults: Findings From Four National Surveys.” Journal of Ger-ontology: Social Sciences 49:S137-45.

Livingston, Ivor Lensworth. 1994. Handbook of Black Americans’Health: The Mosaic of Condi-tions, Issues, and Prospects. Westport, CT: Greenwood.

Lopata, Helena. 1973. Widowhood in an American City. Cambridge, MA: Shenkman.Lugaila, Terry A. 1998. Marital Status and Living Arrangements: March 1998 (update). Wash-

ington, DC: U.S. Department of Commerce.Lund, Dale A., Margaret F. Dimond, Michael S. Caserta, Robert J. Johnson, James L. Poulton,

and J. Richard Connelly. 1985-86. “Identifying Elderly With Coping Difficulties After TwoYears of Bereavement.” Omega 16:213-24.

Matthews, Anne Martin. 1991. Widowhood in Late Life. Toronto: Buttersworth.Mattis, Jacqueline S. and Robert J. Jagers. 2001. “A Relational Framework for the Study of Reli-

giosity and Spirituality in the Lives of African Americans.” Journal of Community Psychol-ogy 29:519-39.

Miech, Richard A. and Michael J. Shanahan. 2000. “Socioeconomic Status and Depression Overthe Life Course.” Journal of Health and Social Behavior 41:162-76.

National Center for Health Statistics. 2000. Health United States, 2000. Hyattsville, MD: U.S.Department of Health and Human Services, Public Health Service, National Institutes ofHealth.

Nelsen, Hart M. and Anne K. Nelsen. 1975. Black Church in the Sixties. Lexington: University ofKentucky Press.

Oggins, Jean, Joseph Veroff, and Douglas Leber. 1993. “Perceptions of Marital InteractionAmong Black and White Newlyweds.” Journal of Personality and Social Psychology65:494-511.

Oliver, Melvin and Thomas Shapiro. 1995. Black Wealth/White Wealth: A New Perspective onRacial Inequality. New York: Routledge.

Orbuch, Terri and Lindsay Custer. 1995. “The Social Context of Married Women’s Work and ItsImpact on Black Husbands and White Husbands.” Journal of Marriage and Family57:333-45.

Orbuch, Terri and Sandra L. Eyster. 1997. “Division of Household Labor Among Black Couplesand White Couples.” Social Forces 76:301-32.

Parkes, Colin M. 1985. “Bereavement.” British Journal of Psychiatry 146:11-17.Parkes, Colin M. and Robert S. Weiss. 1983. Recovery from Bereavement. New York: Basic

Books.Peterson, Christopher. 2000. “The Future of Optimism.” American Psychologist 55:44-55.


Phinney, Jean S. 1996. “When We Talk About American Ethnic Groups, What Do We Mean?”American Psychologist 51:918-27.

Preston, Samuel and Paul Taubman. 1994. “Socioeconomic Differences in Adult Mortality andHealth Status.” Pp. 279-318 in Demography of Aging, edited by Linda G. Martin and SamuelH. Preston. Washington, DC: National Academy Press.

Radloff, Lenore S. 1977. “The CES-D Scale: A Self-Report Depression Scale for Research in theGeneral Population.” Applied Psychological Measurement 1:385-401.

Shaw, Benjamin and Neal Krause. 2001. “Exploring Race Variations in Aging and Personal Con-trol.” Journal of Gerontology: Social Sciences 56B:S119-24.

Shelton, Beth A. and Daphne John. 1996. “The Division of Household Labor.” Annual Review ofSociology 22:299-322.

Singh, Bruce and Beverly Raphael. 1981. “Post-Disaster Morbidity of the Bereaved: A PossibleRole for Preventive Psychiatry?” Journal of Nervous and Mental Disease 169:203-12.

Spanier, Graham. 1976. “Measuring Dyadic Adjustment: New Scales for Assessing the Qualityof Marriage and Similar Dyads.” Journal of Marriage and Family 42:15-27.

Stack, Carol B. and Linda M. Burton. 1993. “Kinscripts.” Journal of Comparative Family Stud-ies 24:157-70.

Stark, Rodney and Roger Finke. 2000. Acts of Faith: Explaining the Human Side of Religion.Berkeley: University of California Press.

Stroebe, Margaret and Wolfgang Stroebe. 1983. “Who Suffers More? Sex Differences in HealthRisks of the Widowed.” Psychological Bulletin 93:297-301.

Sutherland, Lynn, Donna Went, and Elizabeth Douvan. 1990. “Traditionalism in Gender RolesAmong Black and White Newlyweds.” Unpublished manuscript, University of Michigan,Ann Arbor.

Szinovacz, Maximiliane E. 2000. “Changes in Housework After Retirement: A Panel Study.”Journal of Marriage and the Family 62:78-92.

Szinovacz, Maximilane E. and Paula Harpster. 1994. “Couple’s Employment/Retirement Statusand the Division of Household Tasks.” Journal of Gerontology: Social Sciences 49:S125-36.

Taylor, Robert J. and Linda M. Chatters. 1986. “Church-Based Informal Support NetworksAmong Elderly Blacks.” The Gerontologist 26:637-42.

. 1991. “Religious Life of Black Americans.” Pp. 105-23 in Life in Black America, editedby James S. Jackson. Newbury Park, CA: Sage.

Taylor, Robert J., Linda M. Chatters, Belinda J. Tucker, and Edith Lewis. 1991. “Developmentsin Research on Black Families: A Decade in Review.” Journal of Marriage and the Family52:993-1014.

Taylor, Robert J., James S. Jackson, and Linda M. Chatters. 1997. Family Life in Black America.Thousand Oaks, CA: Sage.

Thoits, Peggy. 1995. “Stress, Coping, and Social Support Processes: Where Are We? WhatNext?” Journal of Health and Social Behavior (Extra Issue): 53-79.

Umberson, Debra, Camille B. Wortman, and Ronald C. Kessler. 1992. “Widowhood and Depres-sion: Explaining Long-Term Gender Differences in Vulnerability.” Journal of Health andSocial Behavior 33:10-24.

U.S. Bureau of the Census. 2002. Statistical Abstract of the United States: 2002. 122d ed. Wash-ington, DC: Government Printing Office.

Utz, Rebecca L., Deborah Carr, Randolph M. Nesse, and Camille B. Wortman. 2002. “The Effectof Widowhood on Older Adults’Social Participation: An Evaluation of Activity, Disengage-ment, and Continuity Theories.” The Gerontologist 42:522-33.

Utz, Rebecca, Erin Reidy, Deborah Carr, Randolph M. Nesse, and Camille B. Wortman. 2004.“The Daily Consequences of Widowhood: The Role of Gender and Intergenerational Trans-fers on Subsequent Housework Performance.” Journal of Family Issues 25 (5): 683-712.


Volkart, Edmund H. and Stanley T. Michael. 1957. “Bereavement and Mental Health.” Pp.281-307 in Explorations in Social Psychiatry, edited by Alexander H. Leighton, John A.Clausen, and Robert N. Wilson. London: Tavistock.

White, Lynn K. and Stacy J. Rogers. 2000. “Economic Circumstances and Family Outcomes: AReview of the 1990s.” Journal of Marriage and Family 62:1035-51.

Wickrama, K. A. S., Frederick O. Lorenz, and Rand D. Conger. 1997. “Marital Quality and Phys-ical Illness: A Latent Growth Curve Analysis.” Journal of Marriage and the Family59:143-55.

Williams, David R. and Chiquita Collins. 1995. “U.S. Socioeconomic and Racial Differences inHealth: Patterns and Explanations.” Annual Review of Sociology 21:349-86.

Wilson, William Julius and Kathryn Neckerman. 1986. “Poverty Status and Family Structure:The Widening Gap Between Evidence and Public Policy Issues.” Pp. 232-59 in Fighting Pov-erty: What Works and What Doesn’t, edited by Sheldon Danziger and Daniel H. Weinberg.Cambridge, MA: Harvard University Press.

Zisook, Sidney, Richard DeVaul, and N. Click. 1982. “Measuring Symptoms of Grief andBereavement.” American Journal of Psychiatry 139:1590-93.

Zisook, Sidney, Martin Paulus, Stephen R. Shuchter, and Lewis L. Judd. 1997. “The Many Facesof Depression Following Spousal Bereavement.” Journal of Affective Disorders 45:85-94.

Zisook, Sidney and Stephen R. Shuchter. 1991. “Early Psychological Reaction to the Stress ofWidowhood.” Psychiatry 54:320-32.

Deborah S. Carr, Ph.D., is an assistant professor in the Department of Sociology andInstitute for Health, Health Care Policy & Aging Research at Rutgers University. Herresearch focuses on bereavement and end-of-life issues. She recently received fundingfrom the National Institute on Aging to study influences on end-of-life care prefer-ences.


Documents

Black/White Differences in Psychological Adjustment to Spousal …sites.bu.edu › deborahcarr › files › 2018 › 01 › black_widows.pdf · 2018-01-10 · ARTICLERESEARCH ON